Non-invasive sensing and diagnosis of turbine flow meters

Abstract

Methods and systems for non-invasive sensing and diagnostics of a turbine flow meter may involve: monitoring a turbine flow meter with a non-invasive sensor; extracting a blade frequency associated with the turbine blade based on the flow rate; detecting a pressure wave frequency associated with an internal moving part of the turbine flowmeter; and employing multi-sensor fusion with respect to sensor data generated from the non-invasive sensors including the extracted blade frequency and the pressure wave frequency to identify a health condition of the turbine flowmeter and predict maintenance for the turbine flowmeter. The non-invasive sensor may be mounted in proximity to a position of a moving part of the turbine flow meter, such as a turbine blade, a bearing, a rotor, etc.

Description

Technical Field

[0001] The present invention relates generally to the field of condition monitoring and predictive maintenance in industrial instrumentation. The present invention also relates to methods, systems and apparatus for non-intrusive sensing and diagnostics of turbine flow meters. Background Art

[0002] A turbine meter (also known as a turbine flowmeter) is a device used to measure the flow rate of a fluid, particularly a gas, by utilizing a rotating turbine within a pipe. The rotation of the turbine is proportional to the fluid flow rate, allowing for accurate measurement of volumetric or mass flow rates in a variety of industrial applications.

[0003] Turbine meters have been around since at least 1940, and millions of these mechanical gas flow meters are in use worldwide. For example, a major gas utility in a major city has installed tens of thousands of these meters to measure gas flow across the city gas distribution network for commercial and industrial users. It is estimated that at least 6.5 million turbine meters have been installed worldwide over the past 30 years (1990-2019). However, these devices have mechanical parts that can wear and damage over time, especially when the flow and pressure in the transmission pipeline vary significantly.

[0004] The ongoing shift towards sustainability and decarbonization is driving the increasing adoption of clean, renewable energy sources such as hydrogen (H2) and biogas. These alternative fuels are being injected more frequently into existing gas pipelines. However, hydrogen can negatively impact the fatigue properties of mechanical components, leading to issues such as embrittlement. This, in turn, can cause problems such as cracks and defects in meter rotors, ultimately leading to malfunctions in the blades and bearings.

[0005] Figure 1A 、 Figure 1B 、 Figure 1C and Figure 1D Prior art images 101, 102, 103, and 104, respectively, illustrate damaged meters. Images 101, 102, 103, and 104 depict the condition of the inspected meters, clearly demonstrating mechanical fatigue, which in turn can lead to malfunction of various meter rotor components. Images 101, 102, 103, and 104 show examples of mechanical fatigue evident in these meter mechanisms, a condition that poses a significant risk to the operational integrity of various rotor components within the meters. The visible manifestations of mechanical fatigue during manual and open inspections not only highlight the potential for malfunction within these critical metering systems and the need for remedial measures to mitigate impending problems, but also demonstrate the need for non-intrusive and real-time detection technologies to ensure continued functionality and provide early detection and predictive maintenance planning.

[0006] Figure 2A and Figure 2B An image illustrating a prior art pulse or frequency counter configuration based on the use of an intrusive sensor for a pulse reader. Figure 2A A cutaway perspective view of a meter 120 is shown having a meter housing including circular meter housing end portions 135 and 137 and a pulse pick-off assembly 122. An open vane rotor 126 is located within the meter 120 relative to the stator 124. Figure 2B A cutaway perspective view of the meter 120 is depicted, but with additional components including a downstream stator 132 and a rimmed rotor 129 positioned proximate to the guide ring 127 , as well as an upstream rotor 125 . Figure 2A and Figure 2B The configuration shown in indicates that existing solutions can use direct intrusive pulse picking components for measurement activities. However, such existing applications have the problems discussed above.

[0007] These challenges impact the user experience by potentially producing highly inaccurate or incorrect financial measurement results, compromising the legal metrology and reliability of installed meters. The fact that damaged metal parts and degraded measurement accuracy are not visible from the outside necessitates regular manual inspections. However, identifying (partially) defective equipment through manual inspection is a tedious process that often requires downtime. By the time these issues are discovered, it's often too late. Therefore, there is a clear need for real-time, non-intrusive sensing and diagnostics for existing and new meters to provide customers and users with early visibility into meter status, especially in remote locations, to enable predictive maintenance. Summary of the Invention

[0008] The following summary of the invention is provided to facilitate understanding of some features of the embodiments disclosed herein and is not intended to be a complete description. A comprehensive understanding of the various aspects of the embodiments disclosed herein can be obtained by taking the specification, claims, drawings, and abstract as a whole.

[0009]

[0011] Accordingly, one aspect of the embodiments is to provide methods, systems, and apparatus for non-intrusive sensing and diagnostics of turbine flow meters.

[0010] Another aspect of the embodiments is to provide methods, systems, and apparatus for obtaining, analyzing, and extracting unique characteristics of the condition of a turbine flow meter.

[0011] Another aspect of the embodiments is to provide methods, systems, and apparatus that facilitate automatic detection of anomalies regarding turbine flow meters and provide valuable information for timely and accurate maintenance work.

[0012] The foregoing aspects and other objectives can now be achieved as described herein. In an embodiment, a method for non-intrusive sensing and diagnostics of a turbine flow meter that can be mounted, for example, on a pipeline can include: monitoring the turbine flow meter with a plurality of non-intrusive sensors, the turbine flow meter including turbine blades; extracting blade nominal frequencies associated with the turbine blades based on flow velocity; detecting pressure wave frequencies associated with the turbine flow meter; and employing multi-sensor fusion on sensor data generated from the plurality of non-intrusive sensors, including the extracted blade frequencies and the pressure wave frequencies, including frequency components associated with possible defects, to identify the health of the turbine flow meter and predict maintenance for the turbine flow meter.

[0013] In an embodiment, monitoring a turbine flow meter with a plurality of non-intrusive sensors may further involve capturing acoustic and acceleration signals through walls and a housing of a device under test (DUT) using the plurality of non-intrusive sensors, wherein the DUT comprises a turbine flow meter.

[0014] In an embodiment, extracting blade frequencies associated with turbine blades based on flow rate may involve: determining blade frequencies generated by the turbine blades based on a correlation between a flow rate from a turbine flow meter and a rotational speed of the turbine blades from a unique signal signature extracted from acoustic signals and acceleration signals generated from a plurality of non-intrusive sensors; performing time-frequency analysis on sensor data obtained from the plurality of sensors to detect or eliminate harmonics; and cross-correlating blade-generated frequencies of the turbine blades with sensor-detected frequencies detected by the plurality of non-intrusive sensors.

[0015] In an embodiment, detecting pressure wave frequencies associated with a turbine flow meter may also involve using multiple non-intrusive sensors to detect pressure wave frequencies generated by the rotational effects of turbine blades and potentially defective and / or missing parts in the turbine meter to acquire and evaluate data within a defined two-dimensional moving capture window (MCW).

[0016] In an embodiment, the plurality of non-invasive sensors may include, for example, at least one acoustic sensor and at least one accelerometer.

[0017] In an embodiment, at least one accelerometer may be in contact with a surface coupling of a wall of the turbine flow meter.

[0018] In an embodiment, the at least one accelerometer and the at least one acoustic sensor may generate a frequency response that includes a maximum frequency related to a flow rate that the turbine flow meter is typically capable of measuring.

[0019] In an embodiment, the time-frequency analysis may include high-order moment analysis and bispectrum on the sensor data generated from the plurality of non-invasive sensors to detect and eliminate harmonics.

[0020] In an embodiment, employing multi-sensor fusion with respect to sensor data generated from a plurality of non-intrusive sensors may further involve using multi-sensor fusion to analyze cross-correlation results, extract sensor signatures, and correlate the sensor signatures with normal conditions and abnormal conditions including damaged components of the turbine flow meter.

[0021] Embodiments may also involve issuing reminders and / or alarms to a local or remote location in response to identifying a health condition predicting maintenance for the turbine flow meter to provide early warning for scheduling maintenance and repair of the turbine flow meter.

[0022] In an embodiment, a plurality of non-intrusive sensors may be mounted proximate to the moving parts of the turbine flow meter.

[0023] In an embodiment, a method for non-invasively sensing and diagnosing a turbine flow meter may involve: monitoring the turbine flow meter using multiple non-invasive sensors that can detect the behavior of turbine blades of the turbine flow meter (e.g., internally operating turbine blades); identifying frequencies associated with movement of the turbine blades based on flow rates determined by the multiple non-invasive sensors; detecting pressure wave frequencies associated with operation of the turbine flow meter; performing multi-sensor fusion that combines sensor data from various non-invasive sensors among the multiple non-invasive sensors, wherein the combined sensor data includes both frequencies extracted from the turbine blades and pressure wave frequencies; and analyzing the combined sensor data to determine the health of the turbine flow meter and predict when maintenance of the turbine flow meter is required.

[0024] In an embodiment, a system for non-intrusive sensing and diagnosis of a turbine flowmeter may include: a plurality of non-intrusive sensors for monitoring a turbine flowmeter including turbine blades and associated bearings, wherein the plurality of non-intrusive sensors extract blade frequencies associated with the turbine blades and detect pressure wave frequencies associated with the turbine flowmeter based on flow rate; and a multi-sensor fusion module for performing multi-sensor fusion of sensor data generated from the plurality of non-intrusive sensors including the extracted blade frequencies and the pressure wave frequencies, wherein the sensor data fused by the multi-sensor fusion module facilitates robust and reliable identification of a health condition of the turbine flowmeter and predictive maintenance of the turbine flowmeter.

[0025] In an embodiment of the system, a plurality of non-intrusive sensors may capture acoustic and acceleration signals through the walls and housing of a device under test (DUT), wherein the DUT comprises a turbine flow meter.

[0026] In one embodiment of the system, a nominal blade frequency can be generated by the turbine blades based on a correlation between the flow rate of the turbine flow meter and the rotational speed of the turbine blades driven by the fluid-moving force measured by the turbine flow meter. The true frequency can also be determined from signal features extracted from acoustic and acceleration signals generated by multiple non-intrusive sensors. For example, at a given flow rate, if the turbine flow meter is operating normally, the true frequency determined by the non-intrusive sensors will be within a narrow range of the nominal frequency. On the other hand, the true frequency determined by the non-intrusive sensors may be outside the nominal frequency, indicating the need for further verification of an abnormal condition using cross-correlation and harmonic suppression techniques.

[0027] In an embodiment of the system, time-frequency analysis may be performed on sensor data obtained from a plurality of sensors to detect or eliminate harmonics.

[0028] In an embodiment of the system, the frequencies generated by the blades of the turbine blades may be cross-correlated with the frequencies detected by the sensors detected by the plurality of non-intrusive sensors. Note: See added content in

[018] which is also aligned here.

[0029] In an embodiment of the system, a plurality of non-intrusive sensors may be used to detect pressure wave frequencies generated by blade rotation of turbine blades in a turbine meter within a defined two-dimensional moving capture window.

[0030] In embodiments of the system, the plurality of non-invasive sensors may include one or more acoustic sensors and one or more accelerometers.

[0031] In an embodiment, one or more accelerometers may be in contact with a surface coupling of a wall of the turbine flow meter.

[0032] In an embodiment, one or more accelerometers and one or more acoustic sensors can generate a frequency response that includes a maximum frequency related to the flow rate that the turbine flow meter is typically capable of measuring. Note that the acoustic sensor and accelerometer can be placed in a single unit device or can be separately deployed in two or more unit devices.

[0033] In an embodiment, the time-frequency analysis may include high-order moment analysis and bispectrum on the sensor data generated from the plurality of non-invasive sensors to detect and eliminate harmonics.

[0034] In an embodiment, multi-sensor fusion may be used to analyze the results of the cross-correlation, extract sensor signatures, and correlate the sensor signatures with normal conditions and abnormal conditions including damaged components of the turbine flow meter.

[0035] In an embodiment, in response to identifying a health condition predicting maintenance for the turbine flow meter, reminders and / or alarms may be issued to a local or remote location to provide early warning for scheduling maintenance and repair of the turbine flow meter. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The accompanying drawings also illustrate the present invention and, together with the detailed description of the invention, serve to explain the principles of the invention, wherein like reference numerals refer to identical or functionally similar elements throughout the several views and are incorporated into and form a part of the specification.

[0037] Figure 1A 、 Figure 1B 、 Figure 1C and Figure 1D An image illustrating a prior art damaged meter;

[0038] Figure 2A and Figure 2B A prior art intrusive frequency counter based on a pulse reader is illustrated;

[0039] Figure 3 A flow chart illustrating operation according to an embodiment depicts logical operational steps of a method for non-intrusive sensing and diagnostics of a turbine flow meter;

[0040] Figure 4 A flow chart illustrating operation according to an embodiment depicting logical operational steps of a method for determining blade frequency and revolutions per second (RPS);

[0041] Figure 5 illustrates a table depicting example results obtained from a method for non-intrusive sensing and diagnostics of a turbine flow meter, according to an embodiment;

[0042] Figure 6A and Figure 6B An example of a schematic diagram illustrating an example of a non-invasive sensing device according to an embodiment; and

[0043] Figure 7 A graph depicting processed non-intrusive sensor data indicating an abnormal condition when a blade is missing is illustrated according to an example embodiment.

[0044] In the drawings described herein and shown, identical or similar parts and elements are generally designated by identical reference numerals. DETAILED DESCRIPTION

[0045] The specific values ​​and configurations discussed in these non-limiting examples may vary and are cited merely to illustrate one or more embodiments and are not intended to limit the scope thereof.

[0046] The subject matter will now be described more fully hereinafter with reference to the accompanying drawings, which form a part hereof and show, by way of illustration, specific example embodiments. However, the subject matter is capable of being embodied in a variety of different forms, and thus the subject matter covered or claimed is intended to be construed as not being limited to any example embodiment listed herein; the example embodiments are provided for illustration only. Likewise, the subject matter intended to be claimed or covered is intended to be of appropriately broad scope. Among other things, the subject matter may be embodied as a method, apparatus, component, or system. Thus, embodiments may, for example, take the form of hardware, software, firmware, or a combination thereof. Accordingly, the following detailed description is not intended to be construed in a limiting sense.

[0047] Throughout the specification and claims, in addition to the meanings explicitly stated, terms may have slightly different meanings suggested or implied by the context. Similarly, phrases such as "in one embodiment" or "in an exemplary embodiment" and variations thereof as used herein may not necessarily refer to the same embodiment, and phrases such as "in another embodiment" or "in another exemplary embodiment" and variations thereof as used herein may or may not necessarily refer to different embodiments. For example, claimed subject matter is intended to include, in whole or in part, a combination of the exemplary embodiments.

[0048] In general, terms may be understood at least in part from their usage in context. For example, terms such as "and," "or," or "and / or," as used herein, may include multiple meanings that may depend at least in part on the context in which such terms are used. In general, "or," if used in an associative list, such as A, B, or C, is intended to mean A, B, and C as used herein in an inclusive sense, as well as A, B, or C as used herein in an exclusive sense. Furthermore, the term "one or more," as used herein, may depend at least in part on the context and may be used to describe any feature, structure, or characteristic in a singular sense, or may be used to describe a combination of features, structures, or characteristics in a plural sense. Furthermore, the term "at least one," as used herein, may mean "one or more." For example, "at least one" widget may mean "one or more widgets."

[0049] Terms such as "a," "an," or "the" can also be understood to convey singular or plural usage, depending at least in part on the context. Furthermore, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can allow for the presence of additional factors that are not necessarily explicitly described again, depending at least in part on the context.

[0050] It should be noted that the term "meter" as used herein may refer to a turbine meter or other types of meters related to turbine applications or other rotating applications or rotating equipment such as, for example, compressors. That is, the term "meter" as used herein may refer to a wide range of devices that can be used to measure various quantities in rotating equipment such as, for example, turbine applications or compressors. A "meter" can be used to measure fluid flow rates in rotating equipment such as turbines and compressors, and to monitor other conditions such as, for example, pressure, temperature, and / or other relevant parameters. The term meter therefore may refer to a wide variety of meters employed within turbine systems and other rotating equipment and rotating applications / equipment such as compressors.

[0051] Embodiments relate to methods, systems, and devices for non-intrusive sensing and diagnostics of turbine flow meters. Embodiments can wirelessly monitor turbine flow meters using non-intrusive, battery-powered sensors that capture acoustic and acceleration signals, for example, through the walls and housing of a device under test (DUT). This approach can extract signal features and determine the frequencies generated by turbine blades based on the correlation between flow velocity and turbine rotational speed. Blade rotation can generate disturbances in the turbine meter that form pressure waves, and acoustic / accelerometer sensors can be used to detect the frequencies of the generated pressure waves in a power spectrum by capturing data within a defined, two-dimensional, moving capture window.

[0052] This solution can also perform time-frequency analysis (including high-order moment analysis and bispectral analysis) on the sensor data to detect or eliminate harmonics. The frequencies generated by the blades and the frequencies detected by the sensors, which are associated with the real-time operating flow rate, can be cross-correlated with each other. In addition, multi-sensor fusion can analyze the results of the cross-correlation, extract meaningful sensor features, and correlate the sensor features with normal conditions (e.g., fingerprints and reference files) and abnormal conditions such as damaged or missing components. Predictive maintenance models can generate maintenance plans for turbine flow meters based on the accumulated data to infer the evolution of abnormalities, such as frequency shifts and increased covariance in the time and frequency domains.

[0053] Figure 3 A flowchart illustrating operation according to an embodiment depicts the logical operational steps of a method 200 for non-intrusive sensing and diagnostics of a turbine flow meter. As shown at block 202, a turbine flow meter may be provided, and sensing applications may be performed on the turbine flow meter, such as acceleration detection using an accelerometer, acoustic sensing using one or more acoustic sensors, and surface temperature sensing using one or more temperature sensors. Furthermore, the turbine flow meter may perform flow rate detection using its own flow meter or an additional flow meter.

[0054] Please note that although not Figure 3 Such an example sensor is shown in FIG. Figure 3The various sensing operations are described by one or more steps or operations shown in the various blocks depicted in FIG. For example, as shown at block 204, flow rate sensing steps or operations may be implemented with respect to the turbine flow meter described at block 202 for generating flow rate data associated with the turbine flow meter.

[0055] Similarly, as shown at block 206, one or more steps or operations involving sensors may be implemented with respect to the turbine flow meter shown at block 202. For example, steps or operations as shown at 206 with respect to the turbine flow meter depicted at block 202 may be implemented involving the use of an acceleration sensor for detecting acceleration data, one or more acoustic sensors for detecting acoustic data, and one or more surface temperature sensors for detecting surface temperature data. Such data (e.g., acceleration data, acoustic data, surface temperature data, etc.) may then be cross-correlated with data output as a result of processing the steps or operations shown at block 202 and the following blocks. Figure 4 Details for this cross-correlation operation are shown in block 254 in .

[0056] After processing the flow rate detection operation shown at block 204, a step or operation for correlating the detected flow rate data with a rotational speed or revolutions per second (RPS) may be implemented, as shown at block 205. Thereafter, a step or operation for correlating the blade frequency with the RPS (revolutions per second) may be implemented, as shown at block 207. Note that the step or operation of determining the blade frequency may be verified from processing involving a step or operation of collecting data from, for example, an external frequency counter device as shown at block 201, and there are various ways to verify the frequency associated with the RPS.

[0057] It should be noted that in actual implementations of embodiments based on intrusive measurement methods, an external frequency counter device may not be required. The external frequency counter device can be used for reference verification and / or training for normal conditions during factory testing and can be implemented as an electronic instrument or component that can measure the frequency, including the number of oscillation cycles or pulses per second, in a periodic electronic signal. After processing the steps or operations shown at block 207, steps or operations related to "abnormal conditions" as indicated at block 211 and / or "normal conditions" as shown at block 213 can be implemented.

[0058] It is noted that an abnormal condition may indicate a state or condition that deviates from expected or normal behavior, typically signifying a potential problem or malfunction within the monitored system. On the other hand, a normal condition may represent a standard operating condition or behavior of the system, where everything operates within expected parameters without any anomalies or anomalies. The results of the processed steps or operations, as indicated at block 211 (abnormal condition) and shown at block 213 (normal condition), may be subjected to steps or operations involving generating parameter signatures, as shown at block 215.

[0059] Parametric features can be characteristic patterns or features extracted from the data that can provide insight into the condition or performance of the monitored system and can be used for diagnostic or predictive analysis. Parametric features can include discrete frequency features that can be used as fingerprints and can include other parameters (e.g., Xcorr / cross-correlation data). Validation can be supervised using known cases from model training (block 217). Prediction (block 219) can be implemented as a result of the parametric features and tracking the behavior of frequency shifts and parameters (e.g., covariance trends).

[0060] Returning now to the sensing steps or operations shown at block 206 (i.e., acceleration sensing, acoustic sensing, surface temperature sensing), a method involving a moving part external to the meter but proximate to the interior of the meter mounted on the meter wall as shown at block 208 may be implemented (see, e.g., Figure 7 Next, steps or operations involving acquiring raw data may be implemented, as shown at 210. Note that "raw data acquisition," as used herein, may refer to the process of collecting unprocessed data directly from sensors or other sources without any manipulation or modification.

[0061] Thereafter, as depicted at block 212, operations involving time-frequency analysis (e.g., high-order moment analysis, bispectrum and cross-correlation, singularity identification) may be implemented, followed by steps or operations involving two-dimensional (2D) moving window detection (amplitude and frequency) as shown at block 214. Note that "2D moving window detection" may involve analyzing data within a two-dimensional window that can be moved across the data set to typically identify patterns or anomalies in both amplitude and frequency. Additionally, time-frequency analysis may involve analyzing signals or data in both the time domain and the frequency domain to understand how their characteristics change over time (e.g., frequency fingerprint changes), thereby enabling detection of transient events or frequency changes. Time-frequency analysis may also be performed as a stochastic analysis in the time and frequency domains along with detection including high-order moment analysis (e.g., bispectrum or power spectrum for monochromatic correlated harmonic detection).

[0062] The data generated from the steps or operations shown at block 214 can be implemented in conjunction with the parameter feature operations shown at block 215. Similarly, the data generated as a result of the parameter feature operations depicted at block 215 can be utilized in conjunction with the 2D moving window detection operations shown at block 214. Following the processing of the 2D moving window detection depicted at block 214, a test can be performed, as shown at decision block 221, to determine whether an anomaly has been found. If so, an alert / alarm can be generated, as indicated at block 223, followed by processing of the verification and tuning steps or operations shown at blocks 217 and 215.

[0063] If or when it is determined that no anomaly is found as a result of processing the steps or operations shown at block 221 (anomaly detection test), then the verification steps or operations depicted at block 217 may be implemented to verify that no anomaly is found. Following processing of the verification steps or operations shown at block 217, steps or operations involving prediction may be implemented, as shown at block 219. The results of the verification steps or operations depicted at block 217 may also be included as part of feedback tuning of the parameter feature operations shown at block 215 for further tuning and optimization.

[0064] It is understandable that the above Figure 3 The described method 200 provides several advantages. For example, by utilizing sensors positioned external to the meter but in close proximity to moving parts within the meter, the method 200 enables non-intrusive monitoring with the aid of various external mounting mechanisms such as magnets, threads, and / or adhesive pads. The method minimizes the need for direct contact with the monitored components, thereby reducing the risk of interference with normal operations and simplifying maintenance procedures, particularly for meters that have already been installed. The method 200 can combine multiple sensing modalities, including but not limited to acceleration sensing, acoustic sensing, and surface temperature sensing in addition to flow rate detection. This integrated approach allows for the collection of diverse data types, providing a more comprehensive understanding of the performance and condition of the turbine flow meter.

[0065] Furthermore, by correlating various data streams such as flow rate, rotational speed, and blade frequency, method 200 can facilitate a deeper understanding of the relationships between different operating parameters. This correlation analysis enhances diagnostic capabilities, enabling the detection of abnormal conditions and identification of potential problems at an early stage. Method 200 also involves generating parameter signatures based on the collected data. These parameter signatures can serve as characteristic patterns indicative of the condition of the turbine flowmeter. These parameter signatures can provide valuable information for diagnostics and predictive maintenance, enabling proactive measures to prevent potential failures or malfunctions.

[0066] Through techniques such as time-frequency analysis and 2D moving window detection, method 200 enables the detection of anomalies in the collected data. This allows for the timely identification of deviations from normal behavior, triggering reminders / alarms for further investigation. Additionally, the method supports predictive analysis, which facilitates the prediction of potential problems based on observed patterns and trends. Method 200 also includes steps for verification and validation to ensure the accuracy and reliability of the detection of anomalies. This systematic approach can enhance confidence in the diagnostic results, enabling informed decisions to be made regarding maintenance actions or operational adjustments. Therefore, Figure 3 The method 200 shown in FIG. 2 can provide a robust framework for remote, non-intrusive sensing and diagnostics of turbine flow meters, providing valuable insight into their performance, condition, and identifying potential failure modes.

[0067] Figure 4 An operational flow chart illustrating the logical operational steps of a method 240 for correlating blade frequency with RPS, according to an embodiment, is shown. As shown at block 242, steps or operations may be implemented to determine the turbine meter total volumetric flow rate, Q, expressed in m3 / h. Next, as depicted at block 244, the following calculation may be implemented to determine the RPS in the following equation (Equation 1):

[0068]

[0069] in,

[0070] RPS stands for revolutions per second

[0071] N is the number of blades in the turbine meter under investigation (a constant known from the meter size);

[0072] Pn is the number of pulses per cubic meter (a constant known according to the size and type of meter);

[0073] k is the error correction factor expressed as a percentage (a known constant based on meter size and type);

[0074] D is the diameter of the meter inside the turbine (m);

[0075] v is the flow velocity (m / s)

[0076] Thereafter, as depicted at block 246, steps or operations involving generating RPS data may be implemented, which may then be used as part of steps or operations for determining blade-generated frequencies as shown at block 248 and for determining pressure waves as shown at block 250. The operations shown at block 248 may be implemented based on the following equation (Equation 2):

[0077]

[0078] in,

[0079] f is the frequency output proportional to the flow velocity inside the flowmeter, generated by the rotation of the rotor blades;

[0080] Pn is the number of pulses per cubic meter (a constant known according to the size and type of meter);

[0081] k is the error correction factor expressed as a percentage (a known constant based on meter size and type);

[0082] D is the diameter of the meter inside the turbine (m);

[0083] v is the flow velocity (m / s)

[0084] Following processing of the operation shown at block 250, operations involving sensor fusion may be implemented, as shown at block 252, and operations involving vibration / acoustic frequency detection may be implemented, as depicted at block 256. The data (F (Hz)) obtained from the step or operation shown at block 256 may be used as part of a cross-correlation step or operation, as depicted at block 254, as follows:

[0085] U=Xcorr(f,F) (3)

[0086] in,

[0087] f is the frequency output proportional to the flow velocity inside the flowmeter, generated by the rotation of the rotor blades;

[0088] F is the frequency detected by the acoustic and accelerometer sensors attached to the meter wall surface

[0089] U represents the cross-correlation output

[0090] Likewise, data generated as part of the cross-correlation step or operation shown at box 254 can be used with the vibration / acoustic frequency detection step or operation indicated at box 256. The cross-correlation data determined as a result of the step or operation depicted at box 254 can also be used as part of the sensor fusion operation shown at box 252. The sensor fusion operation shown at box 252 can combine various data provided from acoustic and acceleration sensors (e.g., accelerometers and MEMS microphones) to make a "double check detection" decision. Xcorr (cross correlation) of the bispectral output can be used as one way to fuse the data. However, it should be understood that there are other methods for implementing the disclosed sensor fusion scheme and, for example, image processing.

[0091] Feedback tuning occurs implicitly as part of the implementation of the operations depicted at blocks 248 and 256, and for this feedback coordination, data generated as a result of the sensor fusion step or operation shown at block 252 may likewise be provided to and used as part of the cross-correlation step or operation shown at block 254. Data determined as a result of the step or operation shown at 254 may be used as part of the frequency determination step or operation shown at blocks 248 and 256. Similarly, data from the frequency determination step or operation may be provided to and used as part of the cross-correlation step or operation shown at block 254.

[0092] It is noted that the term "sensor fusion," as used in the context of the steps or operations shown, for example, at block 252, may refer to the process of combining data from multiple sensors to obtain a more comprehensive and accurate understanding of a system or environment than would be achieved by each sensor alone. In the context of the described method 240, sensor fusion may involve integrating data and features from various sensors (such as sensors monitoring vibration, acoustic waves, and other parameters) to more effectively enhance analysis of turbine meter performance and detection capabilities and anomalies and to improve detection confidence levels for some difficult or rare situations.

[0093] Note that block 252 may be implemented as or facilitated by a multi-sensor fusion module that performs the aforementioned multi-sensor fusion of sensor data generated from the non-intrusive sensors discussed herein. The sensor data may include extracted blade frequencies and pressure wave frequencies, and / or other data. Such example sensor data fused by the multi-sensor fusion module may facilitate identification of the health status of the turbine flow meter and predictive maintenance of the turbine flow meter. Block 252 may therefore be referred to as a multi-sensor fusion module.

[0094] It should be noted that the term "module" as used herein can refer to a collection of routines and data structures that perform a specific task or implement a specific data type. A module can be composed of two parts: an interface, which lists the constants, data types, variables, and routines that can be accessed by other modules or routines; and a specific implementation, which can usually be private (accessible only to the module) and includes the source code that actually implements the routines in the module. The term module can also simply refer to an application, such as a computer program designed to help perform a specific task. In some exemplary embodiments, the term "module" can refer to a modular hardware component or a component that is a combination of hardware and software.

[0095] like Figure 4The term "cross-correlation," as used herein with respect to the steps or operations shown at block 254, may refer to a mechanical operation that can be used to measure the similarity between two signals or digital data arrays based on the displacement of one signal or digital data array relative to the other. In the context of method 240, cross-correlation is employed to analyze the relationship between different data streams, such as vibration or acoustic frequencies and blade-generated frequencies. This analysis helps identify patterns or correlations between these signals, providing insight into the operation and condition of the turbine meter.

[0096] The term "blade-generated frequency," as shown, for example, in block 248, may relate to an expression of the frequency at which the blades of a turbine meter rotate. This frequency may be directly related to the rotational speed, or RPS, of the turbine blades and may be critical for determining flow rate through the meter under normal circumstances. In the described method, the blade-generated frequency is calculated based on RPS data obtained from the turbine meter. Analysis of this frequency may be helpful in evaluating the performance and efficiency of the turbine meter and may be used in conjunction with other data, such as vibration or acoustic frequency, for diagnostic purposes and anomaly detection and verification.

[0097] Figure 5 A schematic diagram of a system 260 for non-intrusive sensing and diagnostics of a meter such as a turbine flow meter 261 according to an embodiment is illustrated. The system 260 can integrate various components to achieve efficient monitoring and analysis of meter performance. An important component of the system 260 is a transmitter, such as, for example, a Honeywell Versatilis TM A transmitter (HVT) device 264 that can facilitate two-way wireless communication via Bluetooth Low Energy (BLE) 268. This device can interface with an HVT BLE configurator application 262, allowing configuration and management of the functionality of the HVT device.

[0098] Note: Honeywell Versatilis TM The transmitter is based on BLE for short-range data communication and the latest HVT offers inherently low power, compact design, and quick and easy installation and commissioning. Honeywell Versatilis TM The transmitter can use long distance and low energy consumption The HVT device 264 is an example of a transmitter device that can be implemented according to one or more embodiments. Other types of non-HVT devices can be used in place of HVT-based devices. For example, in some embodiments, the aforementioned sensors can be installed in separate devices rather than a single device.

[0099] Additionally, the HVT device 264 utilizes the LoRa (long range) communication protocol 270 to connect to a LoRa gateway 272. The gateway acts as a bridge to transmit data to WiFi / Ethernet enabled devices, specifically the MIQ-VM 274. Within the MIQ-VM 274, a LoRaWAN provider virtual machine (VM) 276 can manage the communication protocol and interface with the network server and application server, facilitating the transmission of measurement IQ data 278.

[0100] The received measurement IQ data 278 can then be relayed to a diagnostic subsystem 280 for further analysis. This subsystem includes a measurement IQ dashboard 282, accessible for visualization and monitoring via a display / monitor 284. Additionally, a dedicated database 286, potentially hosted or provided by a server 288, can store relevant data for historical analysis and reference. Notably, the described system 260 is adaptable to various types of turbine meters, exemplified here by turbine meter 261 and turbine meter 263. This versatile setup ensures adaptability and scalability to meet diverse metering needs within diverse environments.

[0101] Amidst increasing cost pressures from the privatization of network assets, the demand for remote service solutions is growing. At the same time, utilities are facing significant knowledge loss due to layoffs as experienced operators exit the workforce. The proposed solution addresses these challenges by facilitating real-time data collection from assets for condition-based maintenance (CBM) and presenting health and operational status through optimized dashboards.

[0102] CBM combines newly developed sensors and sensing devices with physics-based, machine learning, and artificial intelligence-based algorithms to detect abnormal changes in equipment or environmental conditions. This proactive approach quickly identifies problems and issues tailored alerts to operations and maintenance, simplifying problem resolution beyond traditional preventive maintenance practices.

[0103] The system operates autonomously, ensuring continuous (e.g., 24 / 7) monitoring to detect equipment failures in their early stages. By continuously comparing processed sensor data characteristics (e.g., temperature, acoustics, and vibration) with a reference database established during installation or calibration, deviations can be flagged to the end user promptly through visual, auditory, and / or electronic notifications.

[0104] Furthermore, in some implementations, recommendation engines can be used to provide end users or service departments with possible root causes and solutions, thereby improving the efficiency of problem resolution. Beyond cost considerations, this approach also improves the safety, reliability, and efficiency of financial measurement. Transitioning from automated to autonomous operations can improve safety and overall business value by enabling systems to autonomously make data-driven decisions related to normal and abnormal operations. This unified understanding of measurement equipment provides actionable intelligence that enables faster detection and diagnosis of problems, increases measurement certainty, prevents downtime, and enhances condition-based monitoring and maintenance to extend calibration intervals.

[0105] The embodiments cover methods, systems and devices for non-invasive monitoring of turbines, using acceleration and acoustic signatures to assess health and detect faults in connected equipment. This proactive approach aims to prevent failures and predict maintenance needs, thereby reducing metering inaccuracies in millions of turbine flow meters due to undetected damage. Additionally, the method provides an alternative flow measurement technique using non-invasive sensors mounted on the outer surface of the flow meter by analyzing the relationship between detected frequencies such as blade rotation (revolutions per second) and flow rate. Importantly, the application range of this method goes beyond turbine flow meters and is applicable to all mechanical equipment with moving parts, such as, for example, compressors.

[0106] The methods and systems described herein consist of four key elements: 1. non-intrusive sensing, in which signals are captured through the walls and housing of the device under test (DUT), eliminating the need for intrusive monitoring techniques; 2. multi-sensor data acquisition involving acoustic and vibration accelerometers that provide time series data within a defined capture window, with the option to include surface and ambient temperature measurements; 3. multi-sensor fusion for analytical detection, which may involve extracting meaningful sensor features, such as sinusoidal and multi-wavelet components, and correlating them with indicators of potential problems such as blade deformation, component damage, or high bearing friction; and 4. predictive maintenance models, in which the accumulated data can be used to develop near-failure mode models, enabling predictive maintenance scheduling to resolve problems before they escalate.

[0107] Figure 6A and Figure 6B A schematic diagram illustrating an example of a non-invasive sensing device according to an embodiment of the present invention is shown. Figure 6A, a system 300 is shown that includes a turbine flow meter 302 and a set of non-intrusive sensors 304, 306 that can be used to monitor a rotor 308 and blades 310, as well as other internal components of the turbine flow meter 302. In some embodiments, each of the non-intrusive sensors 304 and 306 can be configured within the context of a cylindrical housing. Figure 6B In FIG. 3 , an alternative system 301 is shown, which includes Figure 6A , but with the non-intrusive sensors 304 , 306 located in different positions relative to the turbine flow meter 302 . Figure 6A and Figure 6B It is shown that non-intrusive sensors 304 , 306 may be mounted proximate to the moving parts of the turbine flow meter 302 , such as the rotor 308 and blades 310 .

[0108] Figure 7 Graphs 320, 322, and 324, according to example embodiments, depict data on three axes indicating an abnormal condition that may occur when, for example, a turbine rotor blade is missing. Graphs 320 (x-axis), 322 (y-axis), and 324 (z-axis) each depict a processed frequency spectrum with a varying detection wave. Example graphs 320, 322, and 324 depict data indicating that something unusual is occurring with a turbine flow meter, such as turbine flow meter 302, that measures flow velocity relative to a nominal frequency. For example, if a turbine blade is missing or damaged, graphs 320, 322, and 324 may display processed information on one or more axes using a different type of detection wave than would normally occur if the nominal frequency of the flow velocity were closely correlated to the detected wave. Graphs 320, 322, and 324 provide a visual representation of what the non-intrusive sensors 304, 306, for example, detect when an abnormal condition exists, which can help operators identify problems associated with turbine flow meter 302.

[0109] It should be understood that although the operations of the apparatus, system and / or method are shown and described herein in a particular order, the order of the operations may be changed such that certain operations may be performed in a reverse order, or such that certain operations may be performed at least partially simultaneously with other operations. In another embodiment, instructions or sub-operations of different operations may be implemented in an intermittent and / or alternating manner.

[0110] At least some of the operations or features described herein can be implemented using software instructions stored on a computer-usable storage medium for computer execution. For example, one embodiment of a computer program product includes a computer-usable storage medium for storing a computer-readable program.

[0111] A computer-usable or computer-readable storage medium can be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or apparatus or device). Examples of non-transitory computer-usable and computer-readable storage media include semiconductor or solid-state memory, magnetic tape, removable computer disks, random access memory (RAM), read-only memory (ROM), rigid disks, and optical disks. Current examples of optical disks include compact disks with read-only memory (CD-ROM), compact disks with read / write (CD-R / W), digital video disks (DVD), flash memory, and the like.

[0112] Alternatively, the embodiment may be implemented in hardware or in a specific implementation comprising hardware elements and software elements. In an embodiment that does utilize software, the software may include firmware, resident software, microcode, etc.

[0113] In some alternative implementations, the functions indicated in the blocks may occur in the order indicated in the figures. For example, two blocks shown in succession may actually be executed simultaneously, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. It should also be noted that the blocks of the block diagrams and / or flowchart illustrations, and combinations of blocks in the block diagrams and / or flowchart illustrations, may be implemented by a dedicated hardware-based system that performs the specified functions or actions or executes a combination of dedicated hardware and computer instructions.

[0114] Based on the foregoing, it can be appreciated that various embodiments are disclosed herein. For example, in one embodiment, a method for non-intrusive sensing and diagnosis of a turbine flow meter may involve: monitoring the turbine flow meter with multiple non-intrusive sensors, the turbine flow meter including turbine blades; extracting blade frequencies associated with the turbine blades based on flow velocity; detecting pressure wave frequencies associated with the turbine flow meter; and employing multi-sensor fusion on sensor data generated from the multiple non-intrusive sensors, including the extracted blade frequencies and the pressure wave frequencies, to identify the health of the turbine flow meter and predict maintenance requirements for the turbine flow meter.

[0115] In an embodiment, monitoring a turbine flow meter with a plurality of non-intrusive sensors may further involve capturing acoustic and acceleration signals through walls and a housing of a device under test (DUT) using the plurality of non-intrusive sensors, wherein the DUT comprises a turbine flow meter.

[0116] In an embodiment, extracting blade frequencies associated with turbine blades based on flow rate may also involve: determining blade frequencies generated by the turbine blades based on a correlation between a flow rate of a turbine flow meter and a rotational speed of the turbine blades from a signal feature extracted from acoustic signals and acceleration signals generated from a plurality of non-intrusive sensors; performing time-frequency analysis on sensor data obtained from the plurality of sensors to detect or eliminate harmonics; and cross-correlating the blade-generated frequencies of the turbine blades with sensor-detected frequencies detected by the plurality of non-intrusive sensors.

[0117] In an embodiment, detecting pressure wave frequencies associated with a turbine flow meter may also involve: using multiple non-intrusive sensors to detect pressure wave frequencies generated by blade rotation of turbine blades in the turbine meter within a defined two-dimensional motion capture window, wherein the multiple non-intrusive sensors include at least one acoustic sensor and at least one accelerometer.

[0118] In an embodiment, at least one accelerometer may be in contact with a surface coupling of a wall of the turbine flow meter.

[0119] In an embodiment, the at least one accelerometer and the at least one acoustic sensor may generate a frequency response that includes a maximum frequency related to a flow rate that the turbine flow meter is typically capable of measuring.

[0120] In an embodiment, the time-frequency analysis may also include high-order moment analysis and bispectrum on the sensor data generated from the plurality of non-intrusive sensors to detect or eliminate harmonics.

[0121] In an embodiment, employing multi-sensor fusion with respect to sensor data generated from a plurality of non-intrusive sensors may further involve using multi-sensor fusion to analyze cross-correlation results, extract sensor signatures, and correlate the sensor signatures with normal conditions and abnormal conditions including damaged components of the turbine flow meter.

[0122] Embodiments may also involve issuing reminders and / or alarms to a local or remote location in response to identifying a health condition predicting maintenance for the turbine flow meter to provide early warning for scheduling maintenance and repair of the turbine flow meter.

[0123] In an embodiment, a plurality of non-intrusive sensors may be mounted proximate to the moving parts of the turbine flow meter.

[0124] In an embodiment, a method for non-invasively sensing and diagnosing a turbine flow meter may involve: monitoring the turbine flow meter using multiple non-invasive sensors that observe turbine blades of the turbine flow meter; identifying frequencies associated with movement of the turbine blades based on flow rates determined by the multiple non-invasive sensors; detecting pressure wave frequencies associated with operation of the turbine flow meter; performing multi-sensor fusion that combines sensor data from various non-invasive sensors among the multiple non-invasive sensors, wherein the combined sensor data includes both frequencies extracted from the turbine blades and pressure wave frequencies; and analyzing the combined sensor data to determine a health status of the turbine flow meter and predict when maintenance of the turbine flow meter is required.

[0125] In an embodiment, a system for non-intrusive sensing and diagnosis of a turbine flowmeter may include: a plurality of non-intrusive sensors for monitoring a turbine flowmeter including turbine blades, wherein the plurality of non-intrusive sensors extract blade frequencies associated with the turbine blades and detect pressure wave frequencies associated with the turbine flowmeter based on flow rate; and a multi-sensor fusion module for performing multi-sensor fusion of sensor data including the extracted blade frequencies and the pressure wave frequencies generated from the plurality of non-intrusive sensors, wherein the sensor data fused by the multi-sensor fusion module facilitates identification of a health condition of the turbine flowmeter and predictive maintenance of the turbine flowmeter.

[0126] In an embodiment of the system, a plurality of non-intrusive sensors may capture acoustic and acceleration signals through the walls and housing of a device under test (DUT), wherein the DUT comprises a turbine flow meter.

[0127] In an embodiment of the system, a blade frequency that may be generated by a turbine blade based on a correlation between a flow rate from a turbine flow meter and a rotational speed of the turbine blade is determined from a signal feature extracted from acoustic and acceleration signals generated from a plurality of non-intrusive sensors.

[0128] In an embodiment of the system, time-frequency analysis may be performed on sensor data obtained from a plurality of sensors to detect or eliminate harmonics.

[0129] In an embodiment of the system, blade-generated frequencies of the turbine blades may be cross-correlated with sensor-detected frequencies detected by a plurality of non-intrusive sensors.

[0130] In an embodiment of the system, a plurality of non-intrusive sensors may be used to detect pressure wave frequencies generated by blade rotation of turbine blades in a turbine meter within a defined two-dimensional motion capture window, and the plurality of non-intrusive sensors may include at least one acoustic sensor and at least one accelerometer.

[0131] In an embodiment of the system, at least one accelerometer may be in contact with a surface coupling of a wall of the turbine flow meter.

[0132] In an embodiment of the system, the at least one accelerometer and the at least one acoustic sensor may generate a frequency response that includes a maximum frequency related to a flow rate that the turbine flow meter is typically capable of measuring.

[0133] In embodiments of the system, time-frequency analysis may also involve high-order moment analysis and bispectrum on sensor data generated from a plurality of non-intrusive sensors to detect and eliminate harmonics.

[0134] In an embodiment of the system, multi-sensor fusion may be used to analyze the results of the cross-correlation, extract sensor signatures, and correlate the sensor signatures with normal conditions and abnormal conditions including damaged components of the turbine flow meter.

[0135] In embodiments of the system, in response to identifying a health condition predicting maintenance for the turbine flow meter, reminders and / or alarms may be issued to a local or remote location to provide early warning for scheduling maintenance and repair of the turbine flow meter.

[0136] It will be appreciated that the variations and other features and functions disclosed above, or alternatives thereof, may be advantageously combined into many other different systems or applications. It will also be appreciated that various presently unforeseen or unanticipated alternatives, modifications, variations, or improvements may be subsequently made by those skilled in the art, which alternatives, modifications, variations, or improvements are also intended to be encompassed by the following claims.

Claims

1. A method for non-intrusive sensing and diagnosis of a turbine flow meter, comprising: monitoring a turbine flow meter including turbine blades with a plurality of non-intrusive sensors; extracting a blade frequency associated with the turbine blade based on the flow velocity; detecting a pressure wave frequency associated with the turbine flow meter; and Multi-sensor fusion is employed on the sensor data generated from the plurality of non-intrusive sensors, including the extracted blade frequency and the pressure wave frequency, to identify a health condition of the turbine flow meter and predict maintenance for the turbine flow meter.

2. The method of claim 1 , wherein monitoring the turbine flow meter with the plurality of non-intrusive sensors further comprises: Acoustic signals and acceleration signals are captured using the plurality of non-intrusive sensors through walls and a housing of a device under test (DUT), wherein the DUT includes the turbine flow meter.

3. The method of claim 1 , wherein extracting the blade frequency associated with the turbine blade based on the flow rate further comprises: determining the blade frequency generated by the turbine blade based on a correlation between the flow rate of the turbine flow meter and a rotational speed of the turbine blade from signal features extracted from acoustic and acceleration signals generated from the plurality of non-intrusive sensors; performing time-frequency analysis on sensor data obtained from the plurality of sensors to detect or eliminate harmonics; as well as A blade-generated frequency of the turbine blade is cross-correlated with sensor-detected frequencies detected by the plurality of non-intrusive sensors.

4. The method of claim 1 , wherein detecting the pressure wave frequency associated with the turbine flow meter further comprises: The pressure wave frequencies generated by blade rotation of the turbine blades in the turbine meter are detected within a defined two-dimensional motion capture window using the plurality of non-intrusive sensors, wherein the plurality of non-intrusive sensors includes at least one acoustic sensor and at least one accelerometer. 5 . The method of claim 4 , wherein the at least one accelerometer is in contact with a surface coupling of a wall of the turbine flow meter.

6. The method of claim 4, wherein the at least one accelerometer and the at least one acoustic sensor generate a frequency response that includes a maximum frequency related to the flow rate that the turbine flow meter is generally capable of measuring. 7 . The method of claim 1 , wherein the time-frequency analysis further comprises high-order moment analysis and bispectrum on the sensor data generated from the plurality of non-intrusive sensors to detect or eliminate harmonics.

8. The method of claim 1 , wherein employing the multi-sensor fusion with respect to the sensor data generated from the plurality of non-intrusive sensors further comprises: The multi-sensor fusion is used to analyze the results of the cross-correlation, extract sensor features, and correlate the sensor features with normal conditions and abnormal conditions including damage to components of the turbine flow meter.

9. A method for non-invasively sensing and diagnosing a turbine flow meter, the method comprising: monitoring the turbine flow meter using a plurality of non-intrusive sensors observing turbine blades of the turbine flow meter; identifying a frequency associated with movement of the turbine blades based on the flow rate determined by the plurality of non-intrusive sensors; detecting a pressure wave frequency associated with operation of the turbine flow meter; performing multi-sensor fusion that combines sensor data from various non-intrusive sensors among the plurality of non-intrusive sensors, wherein the combined sensor data includes both the frequency extracted from the turbine blade and the pressure wave frequency; and The combined sensor data is analyzed to determine the health of the turbine flow meter and to predict when maintenance on the turbine flow meter is needed.

10. A system for non-intrusive sensing and diagnosis of a turbine flow meter, comprising: a plurality of non-intrusive sensors for monitoring a turbine flow meter including turbine blades, wherein the plurality of non-intrusive sensors extract blade frequencies associated with the turbine blades based on flow velocity and detect pressure wave frequencies associated with the turbine flow meter; and A multi-sensor fusion module is configured to perform multi-sensor fusion of sensor data generated from the plurality of non-intrusive sensors, including the extracted blade frequency and the pressure wave frequency, wherein the sensor data fused by the multi-sensor fusion module facilitates identification of a health condition of the turbine flow meter and predictive maintenance of the turbine flow meter.

Images